Dokumentasi Pengguna Dokumentasi Pengembang | |

Queue Workers

Optional background job processing for backup, update, and rollback operations.

TL;DR

Run php spark queue:work pubvana as a long-running process (via systemd or Supervisor). Without a queue worker, Pubvana falls back to exec() or synchronous execution — which works for most sites. Queue workers are only needed for high-traffic sites or guaranteed async processing.

Details

What the Queue Is For

Pubvana uses the codeigniter4/queue package for three heavy operations that should not block a web request:

  • Backup — creating a full site backup can take 30–120 seconds.
  • Update — downloading and applying a CMS update can take 60–180 seconds.
  • Rollback — restoring from a backup can take 30–120 seconds.

When an admin triggers one of these from the web UI, Pubvana tries three dispatch methods in order:

  1. Queue — if php spark queue:work pubvana is running, the job is pushed to the queue and processed asynchronously in the background. The admin UI polls for progress via SSE.
  2. exec() — if PHP exec() is available (not disabled in php.ini), Pubvana spawns a background Spark command. No queue worker needed.
  3. Synchronous — runs inline with set_time_limit(300) as a last resort. The web request blocks until the operation completes.

Queue Configuration

The queue is configured in app/Config/Queue.php. The default handler is database, which uses the default database connection and stores jobs in the queue_jobs table. Alternative backends that can be configured:

  • Redis — recommended for high-throughput sites. Requires the predis/predis package.
  • Predis — alias for Redis using the Predis library.
  • RabbitMQ — for enterprise message queue setups.

Failed jobs are preserved (keepFailedJobs = true) and can be inspected in the queue_failed_jobs table.

Running the Queue Worker

For development or quick testing:

php spark queue:work pubvana

For production, keep the worker running continuously. Example systemd unit file (/etc/systemd/system/pubvana-queue.service):

[Unit]
Description=Pubvana Queue Worker
After=network.target

[Service]
User=www-data
WorkingDirectory=/var/www/html
ExecStart=/usr/bin/php /var/www/html/spark queue:work pubvana
Restart=always
RestartSec=5

[Install]
WantedBy=multi-user.target

Enable and start:

systemctl enable pubvana-queue
systemctl start pubvana-queue

Alternatively, use Supervisor to manage the process. Add a [program:pubvana-queue] entry to your Supervisor config pointing to php spark queue:work pubvana.

Do I Need a Queue Worker?

For most Pubvana installations: no. The exec() fallback handles async operations without any additional infrastructure. Consider running a queue worker only if:

  • PHP exec() is disabled on your server.
  • You have high traffic and want guaranteed non-blocking admin operations.
  • You want centralised job failure tracking and retry logic.